Curvelet based feature extraction method for breast cancer diagnosis in digital mammogram

This paper proposes a method for breast cancer diagnosis in digital mammogram. The article focuses on using texture analysis based on curvelet transform for the classification of tissues. The most discriminative texture features of regions of interest are extracted. Then, a nearest neighbor classifier based on Euclidian distance is constructed. The obtained results calculated using 5-fold cross validation. The approach consists of two steps, detecting the abnormalities and then classifies the abnormalities into benign and malignant tumors.